Candidate Experience
What a candidate actually goes through, in order.
The session, step by step
- 1
Join
The candidate opens the invite link or enters a room code at /join. No account is required to sit an interview.
- 2
Resume upload
A text-based PDF up to 10 MB. Text is extracted, then parsed into structured skills, projects, and experience. If parsing fails the candidate is told — the interview never proceeds on an unparsed resume.
- 3
Device check
Camera and microphone permissions are requested up front so the session is not interrupted later.
- 4
Adaptive interview
Phases move through resume, technical, system design, and behavioural. Difficulty follows the candidate's performance — strong answers escalate, weak answers get probed.
- 5
Coding round
Problems are solved in an in-browser Monaco editor and executed server-side against test cases. Language, runtime, and output are captured with the submission.
- 6
Wrap-up
Answers are scored across technical depth, communication, architecture, confidence, authenticity, leadership, and reasoning.
Interruptions and connection loss
Session state is persisted continuously. If the connection drops, the candidate resumes at the same question with the same elapsed context rather than restarting. A recruiter can also pause a room, which blocks further input until it is resumed.
What is recorded
Transcripts, code submissions, and behavioural telemetry are stored against the session and scoped to the organisation that owns the room. Resume text sent for analysis has contact details, raw text, and identifiers removed first — the model receives skills, projects, and experience only.